- Boost - lists.boost.org

Re: [boost] Boost Digest, Vol 7345, Issue 1
by Alfredo Correa 25 Sep '24

25 Sep '24

Hi Artyom, Thank you for you for noticing the proposal and the feedback. > It can be excellent to have something like standard MD arrays like numpy is > today for python, but... > yes, this is one of the goals. > 1. You haven't mentioned even once OpenCV which today is the de-facto > standard for numerical computing in today's C++ world. I do scientific computing with large arrays, and nobody uses OpenCV. It would be fair if you had said that it is the de facto standard for image processing, which I don't do. Second, Multi is not a numerical library specifically; it is about the logic and semantics of multidimensional arrays and containers, regardless of the element type. Having said that, the library should still be very performant with numerical data, and for that, I delegate that responsibility to the algorithms it uses behind the scenes, which are very customizable. Unfortunately, people read "arrays" immediately jump to "numerics" and "linear algebra". >From the introduction "The library's primary concern is with the storage and logic structure of data; it doesn't make algebraic or geometric assumptions about the arrays and their elements. (It is still a good building block for implementing mathematical algorithms, such as representing algebraic dense matrices in the 2D case.) In this sense, it is instead a building block to implement algorithms to represent mathematical operations, specifically on numeric data." > While it isn't > perfect - it actually has similar features in terms of array layout - and > what is more important it has a library of high performance algorithms to > process the data. You must talk at least about interoperability between > cv::Mat and this library > Multi is not an algorithms library, but it carefully plays with existing algorithms and facilitates as much as possible to interface with many numerical libraries. > 2. While actual ndarry handling is nice, it is basically a tip of an > iceberg. You need a decent set of high performance algo combined with it. > There is a clear separation of concerns here. The Multi library deals with the data structure, it uses when it needs to fulfill its semantics, the best *existing* algorithms it can in awareness of the datastructure. The Multi library doesn't provide algorithms, it uses the algorithms that are provided to it via different mechanisms. > > Looks like there is a basic BLAS but the typical library has way-way more > to make it useful. > This is an add-on that is separate from the main library. I cannot even guarantee that it will ship with the proposed version of the library. What do you mean by the "typical library" in "the typical library has way-way more to make it useful"? I agree, promising all linear algebra is infinite work, like reimplenting MATLAB or Mathematica, but BLAS has a finite number of functions. The philosophy of the BLAS-adaptor in particular (again, an optional component) is to interface to what BLAS offers and not more. It is more of a model how to interface a legacy library using the features of Multi. > I looked over the code and didn't find include to cblas... Did I miss > something? Also I see request to link cblas/openblas or so. > Yes, you missed that depending on cblas would tie it the application to cblas, not all BLAS implementations can be used through cblas. BLAS is uses through the ABI directly. (basically I have my own version of the cblas header as an implementation detail). Also the BLAS interface is the same for cuBLAS, which is another story. > I did breef look into implementation and it seems to be lacking of > vectorization (use of intrinsics/vector operators) or have I missed? > You missed that this is a generic, not specifically numerical, library. The other thing to take into account is that vectorization/parallelization is still provided by the external algorithms the library uses internally. For example, then dealing with GPU or OpenMP arrays, the library uses thrust algorithms if they are available, which are parallel. vectorization is use by passing execution policies, or whatever mechanism the underlying algorithms use as long as they conform to a certain syntax. (there are some commits coming regarding the passing of execution policies Remember that I don't use many algorithms internally, only those that are necessary to implement the semantics of the array types, assignment, move assignment, construction and destruction. The rest of the operations generate views mostly, so they don't involve computation or algorithms. For the rest, as a user, you have to use algorithms, and I can help if you find that algorithms do not perform as you expect. > 3. Finally I do want to see comparison with stuff like OpenCV, Eigen and > even uBlas (I know it isn't good) > Multi is for multidimensional arrays, not specifically 2D numerical arrays (i.e. matrices). I would be interested in improving the code if someone claims that operations in these libraries are better performant. I wouldn't want the documentation to be a comparison with so many libraries to which there is just very partial overlap. Specially I don't like to compare with frameworks that lock you in into using their own facilities. I consider that Eigen, OpenCV, Kokkos, PETSc are frameworks. Eigen is listed in one column comparing to other libraries. https://gitlab.com/correaa/boost-multi#appendix-comparison-to-other-array-l… I don't fill confident adding an OpenCV column because I don't have experience with it, but feel free to help me adding a library and answering the points of each row in the comparison table. 4. If you do implement many high performance algorithms this library by any > means shouldn't be header only. > I don't implement any high performance algorithms here, I just give the possibility to interface with them and use them internally when needed to fulfill the semantics of a container. > What is not clear to me is why should I use one over some existing solution > like OpenCV? > - Because sometimes your elements are not numerical types. - Because sometimes you want to specify your element types as template parameters not as OpenCV encoded types. - Because sometimes you want arbitrary dimensionality (e.g. 2D, 3D, 6D) to be compile-time. - Because sometimes you want to apply generic algorithms to your arrays (STL, std::sort, std::rotate, std::ranges, boost::algorithms, serialization) - Because sometimes you want to implement function that are oblivious to the actual dimension or your array (e.g. simultaneously view a 3D array of elements as a 2D array of something, for abstraction). - Because sometimes you want to control allocations, use fancy pointers, parameterized memory strategies, polymorphic allocators, - Because sometimes you want precise value semantics - Because sometimes you want to give guarantees in the presence of exceptions (the library doesn't throw exceptions, to be clear). - Because sometimes you want to work with subblocks of an array in the same way to need you work with the whole array. Thanks, Alfredo

4 7

Re: [boost] [Multi] Proposal
by Alfredo Correa 25 Sep '24

25 Sep '24

> > > > Date: Sun, 22 Sep 2024 17:30:04 +0100 > From: John Maddock <jz.maddock(a)googlemail.com> > To: Seth via Boost <boost(a)lists.boost.org> > > >> But, you making numpy-like library... otherwise you wouldn't be > >> interfacing cblas. > > A clear case of diyd/diyd ("Damned If You Do, Damned If You Don't"). Of > all the features you're missing here you decry "At least show ways to > interoperate with existing thirdparty libraries". Now, they went and did > that (if I'm not mistaken, clearly marked as an extension that might not > even be part of the proposed library), and you decry that. There's no > winning. > > +1, I've only been lurking in this conversation, but it seems to me that > trying to compete with low level and highly optimized BLAS libraries is > a fools errand. > I completely agree, I am not even going to try to include vectorized operations inside this library, even as a specialized cases But I will do anything I can for people to be able to exploit them when possible through algorithms and execution policies. At most what I am doing is to interface with BLAS the best I can with some features that are unique to the library in what personally I think is a fairly original way. (capturing a functional-programming interface). Other *numerical* libraries went that way and that is commendable. Thank all for the feedback, Alfredo

1 0

Asset Stewardship Review
by Arnaud Becheler 23 Sep '24

23 Sep '24

I support the C++ Alliance proposal to transfer assets to it and to form a new Steering Committee. I have been using Boost for almost a decade now for my science project, and had some interactions with its community. - My first interaction with the Boost mailing list was to evaluate whether my small science library might be a suitable proposal. The response was pretty clear: no. Boost was seen as too large, and my project too niche (plus, I assume there might have been some quality concerns as well). While I understood the rationale, it did discourage me from further participation in the mailing list. After all, if my project couldn’t be part of Boost, what was the point of staying engaged and spammed ? This experience might help explain at least one data point in the chart from the Alliance proposal, which highlights the mailing list’s decline in participation. - On the other hand, I was still actively using Boost and seeking a community. I was pleasantly surprised by how welcoming the community was to beginners, especially on the Cpplang Slack channel. As a gay person who enjoys sprinkling my Slack posts with rainbow kitties, I never felt ostracized or discriminated against, nor did I witness any bullying. It seems like the community self-regulates quite well, without an apparent need for the enforcement mechanisms proposed by the Foundation. - More significantly, I only became aware of the Foundation through these proposals. I've never had the chance to interact with any of its members, except for Peter, which makes me a bit skeptical of their motivations. In contrast, members of the C++ Alliance—like Vinnie, Rene, and Joaquin—have been consistently accessible and supportive on the Boost channel of the Cpplang Slack. - The quality of the proposal documents is worth noting: the C++ Alliance’s proposal is highly detailed, well-illustrated, and presents a long-term vision that reignited my enthusiasm for the community. In contrast, the Foundation’s proposal felt more like a brief but reassuring affirmation of the importance of our long-held open-source values, without sparking the same level of excitement. Additionally, members of the C++ Alliance were readily available for real-time discussions and debates about their proposal, whereas members of the Foundation were less/not present and communicative on the platforms I personally use. - I believe in values, principles—and results. I can highlight the success of the C++ Alliance with two personal stories. (i) Recently, a colleague mentioned that Boost was outdated and irrelevant. I disagreed and showed them the new website developed by the C++ Alliance. They quickly changed their mind. (ii) Similarly, a colleague who had switched to Rust was convinced C++ was obsolete due to well-known C++ vs. Rust issues. They even compiled a bullet-point list to persuade others to move away from C++. I showed them the C++ Alliance’s Safe C++ initiative in partnership with Sean Baxter. After reading it, they came back excited, saying the proposal addressed every point on their list and could be a game-changer. Unfortunately, I don’t have comparable stories about the Foundation. - As several reviewers have pointed out, there are shared goals and ideologies at the core of the conflict. The community must embrace these common values and reinforce the collaborative spirit that Boost was built on, ensuring the focus remains on innovation and mutual progress rather than division. Continuing down this path of personal conflict only detracts from the mission, diluting the collective potential of the project. It’s essential that we return to a focus on the project's true objectives—technical excellence, inclusivity, and shared success. By prioritizing cooperation and mutual respect, we can ensure that Boost remains a thriving environment for growth. I personally take home these points from the two proposals: - *Foundation Proposal:* - Uphold core values of the open-source community (openness, collaboration, inclusivity). - Push for immediate inclusivity. - Concerned that financial involvement may compromise the project’s integrity. - Recognizes the importance of not becoming too distanced from the community. - Move slowly, but surely, and freely. - *C++ Alliance Proposal:* - Leverage funding to handle unappealing yet necessary tasks. - Focus on modernizing and appealing to younger developers. - Aim to keep the project relevant and attractive to new talent. - Want to maintain a close relationship with developers. - Move fast and throw money at problems. - *Conclusion:* - Both the C++ Alliance and the Foundation share aligned values but focus on different approaches. - They are complementary, addressing different aspects of the same mission. - Their coexistence can provide a balance, acting as counterweights to maintain both ethical standards and practical sustainability. - Transferring assets to the C++ Alliance would enable them to fulfill their part of the mission more quickly. I don't believe it would impede the Foundation's ability to achieve their goals. - The next question for the community is: how can we support the Foundation in return? For example, a reasonably-sized data science project to collect insights on inclusivity and diversity within the Boost community could help them shape future initiatives around the values they prioritize, and would benefit the entire community. Thank you once again for giving us the chance to review these great documents and visions, and for the hard work from all sides that has allowed me to develop my little science project while standing on the towering shoulders of giants ;) Best, Arno

4 3

CODE_OF_CONDUCT.adoc
by Vinnie Falco 23 Sep '24

23 Sep '24

The subject of a code of conduct has come up a few times, and I think it is worth having a discussion. I hope that everyone will find the following statements uncontroversial: 1. Social spaces have rules, either stated or implied 2. Rules need to state clearly what is allowed and what is disallowed, so that people who may be interested in joining can determine if the culture is the right fit. 3. Rules need to be enforced 4. The Boost mailing lists already have rules 5. The rules for the mailing list are outdated 6. The rules for the mailing list are not particularly visible There is value in refreshing the rules for the Boost mailing list, which are currently described here: https://www.boost.org/community/policy.html I think we should do the following: * Update the rules to be relevant for 2024 * Make the rules visible on the sign-up and administration page for the mailing lists (at https://lists.boost.org) * Consider renaming the rules to something other than "Boost Discussion Policy" * Send a periodic announcement to the mailing lists reminding everyone of the rules. These could be timed to coincide with the releases (three times per year) The community needs to discuss the things above and come to a consensus. And we need to make a final decision on what the changes will look like given the feedback. I don't know that the review process is right for this, because it will likely be an iterative process where we all work on a "shared document" to make edits. The nice thing is that we don't need to wait for the outcome of the Boost Asset Stewardship review to do this, because the content of the website and mailing lists is not controlled by the Boost Foundation, the C++ Alliance, or the proposed Steering Committee. And even before the need to decide which revision of the updated rules will be accepted by the project, discussion can take place with no governing body. There is one hiccup though. Where do we discuss this? The mailing list is one possibility. However, anyone who joins the mailing list late, or is not a member of the mailing list, has the problem that the visibility of the discussion is not very great. They would have to use the archives, and the archives are not very pleasant. One cannot see the entire discussion as a single page. And if someone changes the subject, then new related threads of conversation become more difficult to find. And at the end of the day, our workflow requires GitHub issues to triage and assign work items to staff. Therefore, here is a GitHub issue for this. Everyone can feel free to participate in a discussion there. Alternatively, if you feel that the mailing list is a better place, then reply to this post and have a discussion here. There might be two discussions. Or one. Or none (hopefully not). And when we are done we can compare the pros and cons of GItHub issues versus the mailing list. There could be value here. This is the issue: https://github.com/boostorg/website-v2/issues/1269 What do you think? Thanks

3 4

Asset Stewardship Review
by Adam C Getchell 23 Sep '24

23 Sep '24

I support the C++ Alliance proposal to transfer assets and form a new Steering Committee. I’ve been a downstream user of Boost for more than a decade for my research project, and joined the mailing list to get help on using the project effectively. My initial interactions were to sort out build errors, which were mostly resolved with the adoption of vcpkg (after trying biicode and then Conan) and the work on making Boost compatible with CMake. I have appreciated the C++ Alliance’s sponsorship of tools such as the Slack workspace and CI, and I believe this is necessary to ensure progress. I also learned a lot reading the documentation and code of Louis Dionne’s Boost.Hana and Vinnie Falco’s Boost.Beast. As a side note, my research involves writing low-level simulations of the universe, and unfortunately the various memory errors and other bugs have convinced me to go reluctantly down the path of using Rust to write low-level memory-safe abstractions that I can count on to run for months at a time without segfaults or other issues that compromise the validity of the data. Thus, Circle and the recent Safe C++ proposal are very interesting to me, and if they are ever adopted I might back port them to my current codebase. However, this probably won’t be effective until the libraries I depend on make similar efforts. By then, I will probably have everything I need written in Rust. -- Adam Getchell https://adamgetchell.org <https://adamgetchell.org/>

1 0

Boost.LEAF in Embedded Systems (CppCon)
by Vinnie Falco 23 Sep '24

23 Sep '24

At CppCon 2024 the C++ Alliance sponsored a program to interview conference goers about Boost-related topics. This is the first video in the series: Boost.LEAF in Embedded Systems: Insights from Ex-Google Engineer Khalil Estelle at CppCon https://www.youtube.com/shorts/BksBxnTLltY Enjoy! Thanks

1 0

Boost Asset Stewardship Review begins today (3rd Sept)
by Glen Fernandes 22 Sep '24

22 Sep '24

Dear Boost community, The Boost Asset Stewardship Review begins today, 3rd September and ends on 22nd September. For reference, the review schedule is: https://www.boost.org/community/review_schedule.html The submission which initiated the review is the following proposal from the C++ Alliance: https://cppalliance.org/pdf/Fiscal-Sponsorship-Proposal.pdf It proposes a Fiscal Sponsorship which: 1) is a legal agreement where the C++ Alliance holds assets on behalf of the Boost project. It proposes a newly formed (steering) committee that would be composed of Boost developer community members that would determine how the assets are used. 2) donates the assets that it funds to the Boost project (domains, hosting, etc.) The "Boost developer community" here refers to: - Boost library authors - Boost library contributors - Boost release managers For a high level overview of these terms (which I needed) see the following email with definitions: https://lists.boost.org/Archives/boost/2024/09/257579.php The Boost Foundation asks the Boost developer community to consider the following alternative: https://docs.google.com/document/d/1XFt7Bh71e4_uE0iK4jifhR__P0iG5_c1cDfBsMj… The Boost Foundation proposal presents a path forward where it continues to be the steward the assets in question, but offers a change in how these are managed. The assets here refer to the following: - The boost.org domain - Website hosting - Mailing lists - Downloads storage and CDN - Drone CI Currently, the Boost Foundation owns and pays for the following: - Website hosting - Mailing lists The C++ Alliance pays and manages for the following: - The Drone CI - The CDN for the downloads - The download archive - Technical assistance and support for the Boost release process - Backup domain names (in case boost.org expires) - Hosting and downloads for the new website prototype (boost.io) The boost.org domain is absent from both lists above: The Boost Foundation technically does not own it yet. The domain was owned by Beman and he had the credentials. The Boost Foundation is working with Beman's widow and the Dawes Estate to transfer ownership to it. For reference, see Kristen's email: https://lists.boost.org/Archives/boost/2024/08/257563.php Note that the content of the current boost.org site and the new boost.io website is still the province of the Boost developer community. To be clear, the review is not about deciding governance over Boost C++ library development. That remains in the hands of the Boost developer community. Please send all reviews to the Boost developer mailing list. If you do not want to create an account, you can email me at glenfe -at- boost.org with your review and I will post it to the list for you. Include "Asset Stewardship Review" or "Asset Stewardship" in your email subject. In your review do disclose any affiliation to either the C++ Alliance or Boost Foundation groups. Please state any connection you have to Boost (developer, user, package manager etc.) Thank you, Glen

15 26

Boost Asset Stewardship Review concludes 09/22
by Glen Fernandes 21 Sep '24

21 Sep '24

Everyone, Tomorrow (September 22) is the last day to submit a review. To date there have been no requests for an extension. I have received suggestions for potential reviewers I should reach out to, and I believe only one review is forthcoming. Thank you to everyone who has submitted a review, or has otherwise contributed to the discussion. Glen (I was reminded that there are still answers pending from certain members of the Boost Foundation. I understand that CppCon is ongoing and they might not be able to reply yet).

1 0

Re: [boost] Boost Digest, Vol 7346, Issue 1
by Alfredo Correa 20 Sep '24

20 Sep '24

Hi Seth, On Fri, Sep 20, 2024 at 5:00 AM <boost-request(a)lists.boost.org> wrote: > > Date: Thu, 19 Sep 2024 16:28:28 +0200 > From: Seth <bugs(a)sehe.nl> > > On Thu, Sep 19, 2024, at 1:13 PM, Artyom Beilis via Boost wrote: > > 1. You haven't mentioned even once OpenCV > > [...] > > 4. If you do implement many high performance algorithms this library by > any > > means shouldn't be header only. > > > > What is not clear to me is why should I use one over some existing > solution > > like OpenCV? > > All of the points raised are actually a single one, and seem to > misunderstand the scope of the library? I suppose it should be seen as a > successor to Boost MultiArray, leveraging modern language features and > side-stepping some of the pitfalls. > yes, exactly. You seem to understand exactly what I am talking about. (Also a big fan of your Stackoverflow answers). I just gave a long answer to Artyom's message in the list. > That's not at all useless, and interop examples would be very valuable. > However, not all code needs vectorization (beyond what the compiler will > automatically do). I'm more interested in the relation to > https://en.cppreference.com/w/cpp/container/mdspan - which seems to fill > the same space: building block for flexibly indexed contiguous storage. > >From this section: https://gitlab.com/correaa/boost-multi#substitutability-with-standard-vecto… "Multi-dimensinal arrays can interoperate with C++23's non-owning mdspan. Preliminarily, Multi's subarrays (arrays) can be converted (viewed as) mdspan." https://godbolt.org/z/aWW3vzfPj Also see https://gitlab.com/correaa/boost-multi#appendix-comparison-to-other-array-l… Thank you, Alfredo

1 0

Video on Boost.Cobalt available
by Richard 20 Sep '24

20 Sep '24

Utah C++ Programmers has released a video on the Boost.Cobalt coroutines library: Using Coroutines with Boost.Cobalt <https://www.youtube.com/watch?v=yElSdUqEvME> -- "The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline> The Terminals Wiki <http://terminals-wiki.org> The Computer Graphics Museum <http://ComputerGraphicsMuseum.org> Legalize Adulthood! (my blog) <http://LegalizeAdulthood.wordpress.com>

2 1